Fengshan Du

To investigate the forming mechanism of shaped tubes in roll forming of circular tubes, this study conducted roll forming experiments on shaped tubes with different materials and geometries. Using a 30 × 30 mm Q345B square tube as the target, samples were taken from both the planar and corner regions for mechanical property testing and microstructural characterization. The results show that in the uniaxial tensile tests of the planar region, 20Mn2 exhibited the highest tensile strength. Fracture surfaces of all materials displayed evident dimples, second-phase particles, and tear ridges, indicating significant ductile fracture characteristics. In the hardness tests of the corner region, the inner side showed noticeably higher hardness than the outer side, with the maximum value occurring in the inner area. The dislocation density in the inner corner was greater than in the middle and outer regions, a result of metal flow in the corner area. TEM observations of the corner region revealed substantial dislocation pile-ups at grain boundaries, demonstrating severe plastic deformation in this zone.

The final rolling temperature has a significant impact on the grain recrystallization and mechanical properties of rolled materials and is a key factor in the rolling process. With the development of the aerospace industry, higher requirements have been put forward for the quality of TiAl alloys. The suitable rolling temperature range of TiAl alloys is high and narrow, making it difficult to accurately control the final rolling temperature in real-time under the influence of environmental heat transfer and rolling heat. Finite element analysis can simulate the temperature fields, but takes a long time and is not suitable for online monitoring. Neural networks have the characteristic of fast response speeds and can be used for online control and rolling plan optimization. This article proposes a BP neural network prediction model (GABP) based on a genetic algorithm to predict the final rolling temperature. In order to determine the input parameters of the neural network, MATLAB was used to analyze the effects of various factors on the final rolling temperature. The prediction error of GABP is mainly concentrated at 0–1 °C. Compared with fuzzy neural networks (FNN), GABP has a higher prediction accuracy and can effectively predict the final rolling temperature of a TiAl alloy.

The near-isothermal rolling process has the characteristics of multi-variable and strong coupling, and the industrial conditions change constantly during the actual rolling process. It is difficult to consider the influence of various factors in industrial sites using theoretical derivation, and the compensation coefficient is difficult to accurately determine. The neural network model compensates for the difficulty in determining the compensation coefficient of the theoretical model. The neural network can be trained in advance through historical data, the trained network can be applied to industrial sites for prediction, and previous training errors can be compensated for through online learning using real-time data collected on site. But it requires a large amount of effective historical data, so this research uses a combination of production data from a controllable two-roll rolling mill and finite element simulation to provide training data support for the neural network. Five trained neural networks are used for prediction, and the results are compared with industrial site data, verifying the reliability and accuracy of genetic algorithm optimized neural network prediction. We successfully solved the problem of low control accuracy of TiAl alloy outlet thickness during near-isothermal rolling process.